As OpenAI unveils its latest iteration of the GPT large language model, the company finds itself embroiled in a debate that has nothing to do with the artificial intelligence capabilities it develops but everything to do with what it calls them. The new release, dubbed GPT-4o Mini, promises to be a more affordable option for businesses to build their own chatbots compared to its larger counterparts. However, this name—like its predecessors—has sparked widespread criticism for its complexity and lack of flair. Even CEO Sam Altman has come out to acknowledge the need for a new naming convention.
The critique initially surfaced in a public forum when a popular tech YouTuber, Concept Central, took to X-formerly-Twitter to comment on OpenAI’s naming scheme. He humorously pointed out the unwieldy nature of the names, calling for a revamp. Altman’s response to this critique was a rare but welcome admission that the company’s current naming system might indeed need an overhaul. While there isn’t anything inherently wrong with names like “GPT-3.5 Turbo,” the mouthfuls have become a point of contention among users and industry experts alike.
Interestingly, OpenAI seems to have no shortage of creativity when it comes to naming its other projects. Take, for example, the Sora video generator—a name derived from the Japanese word for “Sky” or “Void”—or the DALL-E image generator, which cleverly pays homage to both the Pixar character WALL-E and the surrealist artist Salvador Dalí. Even within the confines of its more secretive projects, the company exhibits a knack for intriguing code names like Q* and Strawberry. The latter is rumored to focus on endowing large language models with advanced reasoning capabilities. Clearly, OpenAI can cook up some pretty engaging names when it wants to.
This contrast begs the question: why does OpenAI stick to the dry and technical “Generative Pre-trained Transformer,” or GPT, for its flagship products? The term is, after all, just a technical classification that could apply to any LLM of this type. One reason might be brand recognition. After all, “ChatGPT” has become nearly synonymous with AI chatbots, and changing the name could risk losing some of the brand equity built up over the years. But as the criticism grows louder, it may be worth considering if the benefits of a rebrand outweigh the risks.
We reached out to OpenAI to see if there are any immediate plans to rename its GPT models, but so far, there has been no official word. Given the monumental success and widespread recognition of ChatGPT and its underlying models, it’s easy to see why the company might be hesitant to change course now. However, as OpenAI continues to innovate and expand its range of products, adopting a more memorable and engaging naming convention could help set the stage for its next generation of technological marvels.
In the grand scheme of things, what’s in a name? For a company leading the charge in artificial intelligence, the answer might be more significant than you think. Whether OpenAI decides to stick with the tried-and-true or venture into more creative territory, one thing is certain: the world will be eagerly watching, and hopefully, not just scratching their heads at the names.